Microsoft adds multi-model AI features to drive Copilot adoption

New Copilot features integrate multiple AI models within a single workflow, enabling cross-model collaboration, improving response accuracy, reducing hallucinations, and allowing users to compare outputs, as Microsoft intensifies efforts to expand AI adoption.

Microsoft has introduced a new set of capabilities for its Copilot research assistant, enabling users to leverage multiple artificial intelligence models within a single workflow. The update marks a significant step in the company’s efforts to enhance the reliability and efficiency of its AI tools while driving wider enterprise adoption.

Multi-model collaboration to improve output quality

At the centre of the update is a new feature called “Critique,” which allows Copilot’s Researcher agent to combine outputs from OpenAI’s GPT and Anthropic’s Claude models. In this setup, GPT generates responses, while Claude evaluates them for accuracy and quality before they are delivered to users. Microsoft said it plans to further evolve this system into a bi-directional workflow, where both models can review and refine each other’s outputs.

The company believes that integrating multiple models will help reduce AI hallucinations—instances where systems produce incorrect or misleading information—while also improving overall productivity. Additionally, Microsoft has introduced a “Model Council” feature, allowing users to compare responses from different AI models side by side, enabling more informed decision-making.

Expanding AI capabilities amid growing competition

The enhancements come as Microsoft continues to expand access to its Copilot ecosystem, including its Copilot Cowork agentic AI tool, which is being rolled out to users in its Frontier programme. The tool, inspired by emerging autonomous AI agents, is designed to support more advanced, task-oriented workflows.

Commenting on the development, Nicole Herskowitz, corporate vice president of Microsoft 365 and Copilot, said in an interview with Reuters, “Having various different models from different vendors in Copilot is highly attractive - but we're taking this to the next level, where customers actually get the benefits of the models working together.”

The move reflects intensifying competition in the AI space, with rivals such as Google advancing offerings like Gemini, while autonomous AI tools continue to gain traction. Microsoft’s latest updates aim to position Copilot as a more robust and reliable platform in the evolving AI landscape.